The problems inherent in using frequency-specific stimuli in the sound field to determine threshold sensitivity are reviewed, including a discussion of some of the specific problems encountered when introducing pure tones, narrow bands of noise, and frequency-modulated (FM) tones. The results of two experiments are reported. In Experiment I, the relationship between pure tones and frequency-modulated tones is developed under earphones in an anechoic chamber, and in two sound-isolated auditory test rooms (not anechoic). Experiment I resulted in the development of a reference threshold sound pressure level for frequency-modulated signals in the sound field. In Experiment II the reference level was applied to a clinical test facility and evaluated with a group of hearing-impaired individuals. The results suggest that the sound-field reference levels accurately reflect monaural threshold under earphones, when the earphone is calibrated to the ANSI, 1969 standard, and the sound field is calibrated to the suggested standard.